- oracledownload.oracle.com/otndocs/tech/semantic_web/pdf/oow2007... why, when, and how to use...

<Insert Picture Here>

Why, When, and How to Use Oracle Database 11g Semantic TechnologiesXavier Lopez, Ph.D, Director, Oracle Server TechnologiesSouripriya Das, Ph.D, Consultant Member, Oracle Server Technologies

Semantic Data Management Characteristics

• Discovery of data relationships across…• Structured data (database, apps, web services)• Unstructured data (email, office documents) Multi-data types

(graphs, spatial, text, sensors)

• Text Mining & Web Mining infrastructure• Terabytes of structured & unstructured data

• Queries are not defined in advance• Schemas are continuously evolving• Overcome isolated systems design• Built on open, industry standards:

• SQL, XML, RDF, OWL, SPARQL

Why is this Useful?

• Designed to represent knowledge in a distributed world

• A method to decompose knowledge into small pieces, with rules about the semantics of those pieces

• RDF data is self-describing; it “means” something

• Allows you to model and integrate DBMS schemas• Allows you to integrate data from different sources

without custom programming

• Supports decentralized data management

• Infer implicit relationships across data

Application Integration

User

Data Ontologies

RDF/OWL

Query & results

Structured & Unstructured Data Sources

Areas of Applicability

• Intelligence, Law Enforcement: • Threat analysis, asset tracking, integrated justice

• Integrated BioInformatics: • Bio-Pathway analysis, protein interaction

• Finance• Fraud detection, Compliance Management

• Web and Social Network Solutions • Recommender, Social Network Analysis, Activity Analysis

• Enterprise Business Applications• Grid resource mgmt, EII, BI, Configuration mgmt.

Integrated Bioinformatics

Source: Siderean Software

Common Information Model (CIM) A Federation of Ontologies

Analytic IntelligenceContext-Specific Access to Knowledge

hasMemberperpetrator

hasMember

alliedWith

perpetrator

hasAlias

hasMember

hasMember

perpetrator

alliedWith

perpetrator

EventPerson

Group

Person

EventPerson

Group

Event

Person

Group

Group

Event

Analytic IntelligenceContext-Specific Access to Knowledge

hasMemberperpetrator

hasMember

alliedWith

perpetrator

hasAlias

hasMember

hasMember

perpetrator

alliedWith

perpetrator

EventPerson

Group

Person

EventPerson

Group

Event

Person

Group

Group

EventW3C RDF & OWL standards

Knowledge Mining Workflows

Information Extraction

Categorization, Feature/term Extraction

Web Resources

News, Email, RSS

Content Mgmt. Systems

Processed Document Collection

Knowledge Mining & Analysis

• Text Indexing using Oracle Text

• Non-Obvious Relationship Discovery

• Pattern Discovery

• Text Mining

• Faceted Search

AnalystBrowsing, Presentation, Reporting, Visualization, Query

SQL/SPARQL Query

Explore

Domain Specific

Knowledge Base

OWL

Ontologies

Ontology Engineering Modeling Process

RDF/OWL

Edit & Transform

• RDF/OWL Data Management

• SQL & SPARQL Query

• Reasoning

• Semantic Rules

• Scalability & Security

• Graph Visualization

• Link Analysis

• Statistical Analysis

• Faceted Search

• Pattern Discovery

• Text Mining

Load, Query

& Inference

Applications &

Analysis

Semantic Data Management Workflow

Other Data Formats

RSS, email

TransactionSystems

Data Sources

Unstructured Content • Entity

Extraction & Transform

• Ontology Engineering

• Categorization

• Custom Scripting

Knowledge Management PlatformAnalyzing Patterns, Trends & Relationships

Entity Matching

Categorization

Non-obvious

Relationship

Analysis

Search and

Text Indexing

Faceted Search

& Discovery

Graph Analysis

Link Analysis

Centrality

Inferencing

(RDFS, OWL)

INFORMATION EXTRACTIONEntity Extraction, Categorization, ETL, Ontology Engineering

PRESENTATION LAYERGraph, Timeline, Metadata, Spatial

Structured DBMS, Unstructured, Spatial, RSS, email, Documents

SPARQL/SQL

APIs

Oracle 11g

RDF/OWL

Knowledge Base

National Security

Financial Risk Analysis

Regulatory Compliance

Life SciencesDrug Discovery

ManufacturingConfiguration Mgmt

Health SciencesBioSurveillance

On

tolo

gy

Ser

vice

s fo

r B

usi

nes

s M

etad

ata

Customer Solutions

Oracle Technologies

Partner Technologies

<Insert Picture Here>

DB Semantic Technology in Oracle 11gTechnical Overview

Semantic Technology: Building Blocks

• Representation• RDF (Resource Description Framework)• Vocabularies

• RDFS (RDF Schema Language)• OWL (Web Ontology Language)

• Inference• Implicit rules that capture semantics of each vocabulary• RDF, RDFS, OWL-Lite, OWL-DL, OWL-Full

• Query• Using graph-patterns in languages such as SPARQL

Semantic Technology Stack

Standards

based

http://www.w3.org/2007/03/layerCake.svg



RDF

• Originally created to encode metadata such as ‘author’, ‘date’, etc. for web resources.

• Recently, it has become popular to relate things in the real-world such as people, places, concepts etc.

• The basic unit of information (fact) is represented as <subject, predicate, object> triple

• Triples together form a graph, connecting pieces of data

subject objectpredicate

:John “32”^^xsd:decimal:age

:Mary:friendOf

Vocabularies (RDFS and OWL)

• RDFS (RDF Schema)• Structuring of resources and properties

• rdfs:class Class of resources

• rdfs:subClassOf hierarchy of classes

• rdfs:subPropertyOf hierarchy of properties

• OWL (Web Ontology Language)• Builds on RDF(S) …

• Property Characteristics: transitivity, symmetry, functional, inverse functional, inverse

• Class construction via set operations and property restrictions • Separate layers have been defined balancing expressibility vs.

implementability: OWL Lite, OWL DL, OWL Full

RDF(S) Example

:John rdf:type :SWcompanyEmployee

Derived Facts

:OracleHQemployee :SWcompanyEmployeerdfs:subClassOf

:John :California

:corpOfficeLocrdf:type

:OracleHQemployee rdfs:subClassOf :SWcompanyEmployee:OracleHQemployee :corpOfficeLoc :California:John rdf:type :Oracle HQ Employee

Asserted Facts <http://example.org/emp>

rdf:type

Graph Pattern Based Query (SPARQL)

SELECT x, yFROM <http://example.org/emp> WHERE { ?x rdf:type ?y }

x y ----------- ------------------------------- :John :OracleHQemployee :John :SWcompanyEmployee


:John :California


rdf:type

OWL Example

:California

:partOf rdf:type owl:TransitiveProperty :California :partOf :USA:USA :partOf :NorthAmerica

:USA :NorthAmerica:partOf :partOf

Asserted Facts (referencing OWL) <http://example.org/gmap>

:partOf owl:TransitivePropertyrdf:type

:California :partOf :NorthAmerica

Derived Facts

:partOf

Advancing W3C Semantic Standards

• Our implementation entirely based on W3C standards (RDF, RDFS, OWL)• Native SPARQL support is planned

• Members of following W3C Web Semantic Activities:• W3C Data Access Working Group (DAWG)• W3C OWL 1.1 Working group• W3C Semantic Web Education & Outreach (SWEO)• W3C Health Care & Life Sciences Interest Group (HCLS) • W3C Multimedia Semantics Incubator group

22

Technical Features

• Storage model, loading, and management for data represented in RDF/OWL

• SQL-based query of RDF/OWL data• Ontology-assisted query of Relational data

• Native inferencing engine to infer new relationships from RDF/OWL data

Functionality: Overview

RDF/OWL data and

ontologies & rulebases

Enterprise (Relational)

data

Query RDF/OWL data

and ontologies

INFERS

TO

RE

QUERY

RD

F/S

Use

r-de

f.

Bat

ch-

Load

Incr

. D

ML


RDF/OWL data and



data

Query RDF/OWL data

and ontologies

INFERS

TO

RE

QUERY

RD

F/S

Use

r-de

f.

Bat

ch-

Load

Bul

k-Lo

ad

Incr

. D

ML


RDF/OWL data and



data

Query RDF/OWL data

and ontologies

INFERS

TO

RE

QUERY

RD

F/S

Use

r-de

f.

Bat

ch-

Load

OW

Lsub

sets

Bul

k-Lo

ad

Incr

. D

ML


RDF/OWL data and



data

Query RDF/OWL data

and ontologies

INFERS

TO

RE

Ontology-Assisted Query of

Enterprise Data

QUERY

RD

F/S

Use

r-de

f.

Bat

ch-

Load

OW

Lsub

sets

Bul

k-Lo

ad

Incr

. D

ML

Storage: Overview

Model 1

Model 2

Model n

…

OWL subset

RDF / RDFS

Rulebase m

…

Inferred Triple Set 1


Inferred Triple Set p

A1

A2

An

Application Tables with RDF object

type columns

…

RDF/OWL data

and ontologies

Vocabularies and Rulebases

Rules Indexes

(Derived data)R

R

R

her

man

sco

ttsc

ott

Storage: Overview

Model 1

Model 2

Model n

…

OWL subset

RDF / RDFS

Rulebase m

…



Inferred Triple Set p

A1

A2

An

Application Tables with RDF object

type columns

…

RDF/OWL data

and ontologies

Vocabularies and Rulebases

Rules Indexes

(Derived data)R

R

R

Oracle DB Semantic Network (inside MDSYS)

her

man

sco

ttsc

ott

Major Steps for building semantic app.


• Create Semantic Network in Oracle


• Create Semantic Network in Oracle• Create an RDF/OWL model associating with a table column of

type SDO_RDF_TRIPLE_S




• Store new RDF/OWL data into Oracle via bulk-load, batch-load, or SQL INSERT; Also, perform any DML





• Optionally, create user-defined rulebases





• Optionally, create user-defined rulebases• Infer new RDF/OWL data using Oracle’s native inference engine

(OWLPRIME, OWLSIF, RDFS++, RDFS & User-defined Rules)







• Query RDF/OWL data and ontologies via SQL using SEM_MATCH table function; optionally combine with SQL operations such as join, order by, group by, filter conditions, etc.







• Query RDF/OWL data and ontologies via SQL using SEM_MATCH table function; optionally combine with SQL operations such as join, order by, group by, filter conditions, etc.

• Perform Ontology-assisted Query against enterprise (relational) data using SEM_RELATED and SEM_DISTANCE operators

Creating Semantic Network and Semantic Models


Creating a semantic network1. SEM_APIS.CREATE_SEM_NETWORK (<tablespace>);


Creating a semantic network1. SEM_APIS.CREATE_SEM_NETWORK (<tablespace>);

Creating a semantic model2. Create an application table with an SDO_RDF_TRIPLE_S type col

CREATE TABLE ATAB (ID int, TRI SDO_RDF_TRIPLE_S) compress;

3. Create a model associated with the SDO_RDF_TRIPLE_S column

SEM_APIS.CREATE_SEM_MODEL (‘MODEL1’, -- <model_name> ‘ATAB’, -- <app_table_name> ‘TRI’ -- <RDF type col name>

);

Access Control

Access Control

• Models• A database view owned by MDSYS gets created at model creation• The creator gets SELECT privilege with GRANT option• DML on a model is done via DML on the associated RDF object type

column and requires invoker to have appropriate privileges on the associated application table

Access Control



• Rulebases• A database view owned by MDSYS gets created at rulebase creation• The creator gets SELECT and DML privilege with GRANT option

Access Control



• Rulebases• A database view owned by MDSYS gets created at rulebase creation• The creator gets SELECT and DML privilege with GRANT option

• Rules Indexes (Inferred Triple Sets)• A database view owned by MDSYS gets created at rules index creation• Creator must have SELECT privilege on underlying model and rulebase

views• The creator gets SELECT privilege with GRANT option

Storage: Highlights

Storage: Highlights

• Stores <subject, predicate, object> triples• Removes duplicates to ensure RDF/OWL graph is a set• Uses RDF-specific compression for tables and indexes

Storage: Highlights


• No limits on amount of data that can be stored• Current users: Swissprot: 800million triples, UTH: 600million+

Storage: Highlights



• Can handle multiple lexical forms of the same value• Ex: “00123”^^xsd:decimal and “123”^^xsd:decimal• Ex: “2004-12-21T22:00:00-08:00”^^xsd:dateTime and

“2004-12-22T01:00:00-05:00”^^xsd:dateTime

Storage: Highlights




“2004-12-22T01:00:00-05:00”^^xsd:dateTime

• Maintains fidelity (user-specified lexical form)

Storage: Highlights




“2004-12-22T01:00:00-05:00”^^xsd:dateTime

• Maintains fidelity (user-specified lexical form)• Provides access control for models, rulebases, and rules indexes

Storage: Highlights




“2004-12-22T01:00:00-05:00”^^xsd:dateTime

• Maintains fidelity (user-specified lexical form)• Provides access control for models, rulebases, and rules indexes

• Supports long literal values

Loading RDF/OWL data


Load data using


Load data using

• Bulk-load (very fast)• Load data into a staging table (using SQL*Loader from a file

or Named Pipe containing N-Triple formatted data)• Invoke PL/SQL API to invoke bulk load from the staging table


Load data using



• Batch-load (fast, can handle long literals as well)• Invoke Java-based API to load from file containing N-Triple

formatted data


Load data using



• Batch-load (fast, can handle long literals as well)• Invoke Java-based API to load from file containing N-Triple

formatted data

• SQL INSERT (for loading small amounts of data)

Loading APIs: Bulk-Load


• Use SQL*Loader to load staging table• Control file template is available in 11g companion CD

• Only the Staging Table name may need to be changed• Staging Table definition is shown in documentation

• Use COMPRESS option, if available• Input file must be N-Triple formatted

• Named Pipe may be used to save disk space


• Use SQL*Loader to load staging table• Control file template is available in 11g companion CD

• Only the Staging Table name may need to be changed• Staging Table definition is shown in documentation

• Use COMPRESS option, if available• Input file must be N-Triple formatted

• Named Pipe may be used to save disk space

• SEM_APIS.BULK_LOAD_FROM_STAGING_TABLE• Model_owner• Table_owner• Table_name• Flags (default NULL): ‘ VALUES_TABLE_INDEX_REBUILD ‘

Loading APIs: Batch-Load

• Batch-load uses the oracle.spatial.rdf.client.BatchLoader class packaged in <ORACLE_HOME>/md/jlib/sdordf.jar

• Example (on Linux)• java

-Ddb.user=scott -Ddb.password=password -Ddb.host=127.0.0.1 -Ddb.port=1522 -Ddb.sid=orcl -classpath ${ORACLE_HOME}/md/jlib/sdordf.jar:${ORACLE_HOME}/jdbc/lib/ojdbc5.jar oracle.spatial.rdf.client.BatchLoader <N-TripleFile> <tablename> <tablespaceName> <modelName>

• SPARQL-like graph pattern embedded in SQL query• Matches RDF/OWL graph patterns with patterns in stored data

• Returns a table of results

• Can use SQL operators/functions to process results• Avoids staging when combined with queries on relational data

SELECT t.x …

FROM …, TABLE (SEM_MATCH invocation

) t, …

WHERE …

Query RDF Data




SELECT t.x …


) t, …

WHERE …

Query RDF Data

SEM_MATCH ( '(?x :partOf :NorthAmerica)', -- pattern: all parts of N.A. SEM_Models(‘gmap'), -- RDF/OWL data models

SEM_Rulebases(‘OWLPRIME'), -- rulebases SEM_Aliases(…) -- aliases null -- no filter condition)




SELECT t.x …


) t, …

WHERE …

Query RDF Data

Table Columns returned by SEM_MATCH

Each returned row contains one (or more) of the following cols for eachvariable ?x in the graph-pattern:

Column Name Type Description

x varchar2 Value matched with ?x

x$rdfVTYP varchar2 Value TYPe: URI, Literal, or Blank Node

x$rdfLTYP varchar2 Literal TYPe: e.g., xsd:integer

x$rdfCLOB CLOB CLOB value matched with ?x

x$rdfLANG varchar2 LANGuage tag: e.g., “en-us”

Projection Optimization: Only the columns referred to by the containing query are returned.

RDF Query in SQL


:John :California


:USA :NorthAmerica:partOf :partOf

SELECT e, p FROM TABLE(SEM_MATCH( ‘(?e rdf:type ?empCategory) (?empCategory :corpOfficeLoc ?loc) (?loc :partOf ?p)’, SEM_Models(‘emp’, ‘gmap’), SEM_Rulebases(‘OWLPRIME’), SEM_ALIASES(SEM_ALIAS('', 'http://www.example.org/')),NULL));

:partOf

E P

:John :USA

:John :NorthAmerica

Ontology-Assisted Query: Overview

• Motivation• Traditionally relationship between two terms is checked only

in a syntactic manner• Need a new operator which can do semantic relationship

check by consulting an ontology

• Introduces two operators• SEM_RELATED (<col>,<pred>, <ontologyTerm>,

<ontologyName> [,<invoc_id>])• SEM_DISTANCE (<invoc_id>) Ancillary Oper.

Example: Query using Semantic Operators

Finger_Fracture

Arm_Fracture

Upper_Extremity_Fracture

Hand_FractureElbow_FractureForearm_Fracture

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

ID DIAGNOSIS

1 Hand_Fracture

2 Rheumatoid_Arthritis

Patients

“Find all entries in diagnosis column that are related to ‘Upper_Extremity_Fracture’”

Syntactic query will not work:SELECT p_id, diagnosis FROMPatients WHERE diagnosis = ‘Upper_Extremity_Fracture;


Finger_Fracture

Arm_Fracture



rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

ID DIAGNOSIS

1 Hand_Fracture


Patients



SELECT p_id, diagnosis FROM PatientsWHERE SEM_RELATED (diagnosis, ‘rdfs:subClassOf’,‘Upper_Extremity_Fracture’, sem_models(‘NCI’), sem_rulebases(‘OWLPRIME’), …) = 1;


Finger_Fracture

Arm_Fracture



rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

ID DIAGNOSIS

1 Hand_Fracture


Patients




Finger_Fracture

Arm_Fracture



rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

ID DIAGNOSIS

1 Hand_Fracture


Patients



SELECT p_id, diagnosis FROM PatientsWHERE SEM_RELATED (diagnosis, ‘rdfs:subClassOf’,‘Upper_Extremity_Fracture’, sem_models(‘NCI’), sem_rulebases(‘OWLPRIME’),… , 777) = 1 AND SEM_DISTANCE(777) <= 2;


Finger_Fracture

Arm_Fracture



rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

rdfs:subClassOf

ID DIAGNOSIS

1 Hand_Fracture


Patients



Inference: Overview

• Native inferencing for• RDF, RDFS• OWL subsets• User-defined rules

• Rules are stored in rulebases

• RDF/OWL graph is entailed (new triples are inferred) by applying rules in rulebase(s) to model(s)

• Inferencing is based on forward chaining: new triples are inferred and stored ahead of query time

Inferencing

• RDFS Example• rdfs:subClassOf is transitive• rdfs:subPropertyOf is transitive• X rdf:type C, C rdfs:subClassOf SC => X rdf:type SC

• OWL Example• :partOf rdf:type owl:TransitiveProperty• :friendOf rdf:type owl:SymmetricProperty

• User-defined Rule Example:• X :hasParent Y, Y :hasBrother Z => X :hasUncle Z

OWL Subsets Supported

• RDFS++• RDFS plus owl:sameAs and owl:InverseFunctionalProperty

• OWLSIF (OWL with IF semantics)• Based on Dr. Horst’s pD* vocabulary¹

• OWLPrime• rdfs:subClassOf, subPropertyOf, domain, range

• owl:TransitiveProperty, SymmetricProperty, FunctionalProperty, InverseFunctionalProperty, inverseOf

• owl:sameAs, differentFrom

• owl:disjointWith, complementOf,

• owl:hasValue, allValuesFrom, someValuesFrom• owl:equivalentClass, equivalentProperty

• Jointly determined with domain experts, customers and partners

1 Completeness, decidability and complexity of entailment for RDF Schema and a semantic extension involving the OWL vocabulary

OWL DLOWL Lite

OWLPrime

40

11g OWL Inference PL/SQL API

• SEM_APIS.CREATE_ENTAILMENT(• Index_name• sem_models(‘GraphTBox’, ‘GraphABox’, …), • sem_rulebases(‘OWLPrime’),• passes,• Inf_components,• Options)• Use “PROOF=T” to generate inference proof

• SEM_APIS.VALIDATE_ENTAILMENT(• sem_models((‘GraphTBox’, ‘GraphABox’, …), • sem_rulebases(‘OWLPrime’),• Criteria,• Max_conflicts,• Options)

• Above APIs can be invoked from Java clients through JDBC

Typical Usage:

• First load RDF/OWL data

• Call create_entailment to generate inferred graph

• Query both original graph and inferred data

Inferred graph contains only new triples! Saves time & resources

Typical Usage:

• First load RDF/OWL data

• Call create_entailment to generate inferred graph

• Call validate_entailment to find inconsistencies

41

Performance Evaluationusing Desktop

42

Database Setup

• Linux based commodity PC (1 CPU, 3GHz, 2GB RAM) ¹• Database installed on machine “semperf3”

semperf3 semperf1

semperf2

Giga-bit Network

Database 11g

• Two other PCs are just serving storage over network

43

1 http://ontolog.cim3.net/file/work/DatabaseAndOntology/2007-10-18_AlanWu/RDBMS-RDFS-OWL-InferenceEngine--AlanWu_20071018.pdf

Oracle 11g Bulk-Loader Performance

Oracle 11g Bulk-Loader PerformanceLUBM-50 (6.9 million triples)

Oracle 11g Bulk-Loader PerformanceLUBM-50 (6.9 million triples)• TIME 0 hour 13.5 min (31 mil / hr)


• SQL loader: 4 min 56 sec


• SQL loader: 4 min 56 sec • bulk_load API: 8 min 33 sec



LUBM-500 (69 million triples)



LUBM-500 (69 million triples)• TIME 3 hour 20 min (21 mil / hr)




• SQL loader: 0 hour 46 min




• SQL loader: 0 hour 46 min• bulk_load API: 2 hour 34 min





LUBM-1000 (138 million triples)






• SQL loader: 1 hour 34 min







• DISK SPACE 19 GB (7.3 mil / GB)








• Triples and Values: 5 GB








• Triples and Values: 5 GB• Indexes: 11 GB








• Triples and Values: 5 GB• Indexes: 11 GB• App Table (compressed): 3 GB

Query Performance

Ontology LUBM50

6.8 million & 3+ million inferred

LUBM Benchmark Queries

Q1 Q2 Q3 Q4 Q5 Q6 Q7

OWLPrime

# answers 4 130 6 34 719 393730 59

Time

(sec)0.09 0.80 0.28 6.1 0.4 36.82 1.05

OWLPrime +

Pellet on TBox

# answers 4 130 6 34 719 519842 67

Time

(sec)0.09 0.79 0.28 9.5 0.4 43.65 1.13

Query Performance (2)

Ontology LUBM50

6.8 million & 3+ million inferred

LUBM Benchmark Queries

Q8 Q9 Q10 Q11 Q12 Q13 Q14

OWLPrime

# answers 5916 6538 0 224 0 228 393730

Time

(sec)2.76 2.08 0.18 0.02 0.01 0.39 31.5

OWLPrime +

Pellet on TBox

# answers 7790 13639 4 224 15 228 393730

Time

(sec)3.9 3.47 0.37 0.02 0.03 0.39 39.4

Inference PerformanceOntology (size)(after duplicate elimination)

RDFS OWLPrime

#Triples inferred

(millions)

Time(speed)

#Triples inferred

(millions)

Time(speed)

LUBM-506.6 million

2.75 12.25 min(13.5 mil / hr)

3.05 8 min(22.9 mil /

hr)

LUBM-1000133.6 million

55.09 7 hr 19 min(7.5 mil / hr)

61.25 7 hr(8.75 mil /

hr)

UniProt20 million

3.4 24.1 min(8.5 mil / hr)

50.8 3 hr 1 min(16.8 mil /

hr)

Results collected on a single CPU PC (3GHz), 2GB RAM (1.4G dedicate to DB), Multiple Disks over NFS

48

Oracle 11g RDF/OWL

• New features• Bulk loader• Native OWL inference support (with optional proof generation)• Semantic operators

• Performance improvement• Much faster compared to 10gR2

• Loading• Query (table function re-write, query hints)• Inference

• Shipped in 2007

• Java API support• Oracle Jena adaptor (released on OTN)• Sesame (forthcoming)

49

For More Information

search.oracle.com

or

oracle.com

semantic technologies